COF-C03 domain - 10% of the exam

Data Collaboration

Data Collaboration is 10% of the SnowPro Core Certification (COF-C03) exam. These are the objectives it covers, each with practice questions and worked explanations.

Objectives in this domain

Sample question from this domain

Free sampleData Collaborationmedium

A data engineer runs CREATE TABLE sales_clone CLONE sales on a large permanent table and then inserts and deletes rows in the clone over the following week. Which statements correctly describe how this zero-copy clone behaves? Select TWO.

CREATE TABLE sales_clone CLONE sales;
  • AAt the moment of creation the clone references the same micro-partitions as the source, so it adds no extra storage until later DML on either table writes new partitions. Correct
  • BWrites to the clone are reflected back in the source table as well, because both objects continue to point at one shared and continually synchronised set of micro-partitions.
  • CThe clone is independent after creation, so an INSERT or DELETE on the clone writes new micro-partitions for the clone alone and leaves the source table's data unchanged. Correct
  • DThe clone inherits the source table's load history, so a COPY INTO that already loaded a file into the source will skip that same file when run against the clone.
A zero-copy clone shares source micro-partitions at creation, costs no extra storage initially, and is an independent writable object thereafter. Zero-copy cloning creates a new object that points at the source table's existing immutable micro-partitions, so it adds no storage when created; once it exists the clone is fully independent, and any INSERT, UPDATE, or DELETE writes new micro-partitions for that object alone without touching the source, while metadata such as load history is not inherited.

Why A is correct: Cloning copies only metadata pointers to the source micro-partitions, so creation consumes no additional storage and divergent storage accrues only as either table is modified.

Why B is wrong: This is tempting because the clone starts by sharing partitions, but the clone is fully independent after creation, so DML on it writes its own partitions and never alters the source.

Why C is correct: A clone is a separate writable object, so subsequent DML diverges its partition set from the source and modifications never propagate back to the original table.

Why D is wrong: Load history is metadata that is not carried into a clone, so the same staged files can be loaded again into the clone rather than being skipped as duplicates.

Other domains in this exam

See also the COF-C03 cert hub, the study guide, and the cheat sheet.

Examworthy is not affiliated with or endorsed by Snowflake. Original, blueprint-aligned practice material only.